With the widespread use of autonomous systems, liability often shifts towards the manufacturers and part suppliers, especially when various system components, sourced from different suppliers, contribute to a failure. This necessitates a framework for automatic liability apportionment which can be embedded in manufacturer-supplier contracts and minimise legal disputes. To this end, we propose a formal framework based on the notion of actual causality in structural causal models and the robustness semantics of logical specifications. We prove several desirable properties of this framework. Moreover, we formalise the notion of causal non-interaction, give sufficient conditions for it to hold, and demonstrate its utility in deriving upper bounds for liability analysis. Furthermore, we relate our definition to the existing notion of harm, empirically evaluate our framework to demonstrate its efficacy, and release a software package implementing our approach. We extend our framework to handle interval specifications.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Causal Liability in Autonomous Systems

  • Kaveh Aryan,
  • Hana Chockler,
  • Mohammad Reza Mousavi

摘要

With the widespread use of autonomous systems, liability often shifts towards the manufacturers and part suppliers, especially when various system components, sourced from different suppliers, contribute to a failure. This necessitates a framework for automatic liability apportionment which can be embedded in manufacturer-supplier contracts and minimise legal disputes. To this end, we propose a formal framework based on the notion of actual causality in structural causal models and the robustness semantics of logical specifications. We prove several desirable properties of this framework. Moreover, we formalise the notion of causal non-interaction, give sufficient conditions for it to hold, and demonstrate its utility in deriving upper bounds for liability analysis. Furthermore, we relate our definition to the existing notion of harm, empirically evaluate our framework to demonstrate its efficacy, and release a software package implementing our approach. We extend our framework to handle interval specifications.